ethics in policing administration

JAKE MONAGHAN Boundary Policing

I. INTRODUCTION

The structure of institutions—including the boundaries of their authority— can be as significant for securing just and legitimate outcomes as the rules that govern their operations.1 For example, the separation of executive, legislative, and judiciary powers is widely regarded as necessary for good governance; the executive ought not to make the rules and enforce them. Without such institutional boundaries, the enforcement of rules is likely to become arbitrary or discriminatory, thus illegitimate, and the entity enforcing the rules is at risk of becoming illegitimate as well.

Executive power is subject to a number of additional boundaries that are well entrenched yet more fine-grained. The powers of national defense are separated from the powers of domestic law enforcement. The power of making an arrest is distinct from the power to punish. Police are them- selves subject to a variety of geographic and statutory boundaries. Federal agencies, for example, are much more specialized than city police depart- ments or county sheriff’s offices. Unlike the most foundational boundaries of political power, the organization of lower-level institutions like policing has often arisen in practice without theoretical examination. I’ll argue that the political power in our institutions of policing is poorly allocated. This paper aims to remedy that by developing a theoretical framework for eval- uating institutional structures of policing.

Thanks to Ryan Muldoon, Kirun Sankaran, J. P. Messina, Danny Shahar, and Bill Glod for valuable discussion of this article. Thanks also to an audience at the 2020 Philosophy, Politics, and Economics Society Meeting for feedback on a presentation of these arguments. I am especially grateful to the editors at Philosophy and Public Affairs for insightful comments and criticism that significantly improved the article.

1. I rely on a normative conception of legitimacy in this paper, following Buchanan’s use of the term: “an entity has political legitimacy if and only if it is morally justified in wielding political power.” Allen Buchanan, “Political Legitimacy and Democracy,” Ethics 112, no. 4 (July 2002): 689.

© 2020 Wiley Periodicals LLC Philosophy & Public Affairs 49, no. 1

Normative questions about the form of policing have recently taken center stage in political discourse as protests against police brutality have swept the United States.2 Increasingly, the possibility of reorganizing, defunding, or even abolishing police institutions is being taken seriously. These responses to the ills of American policing are not focused simply on the behavior of police officers or the justice or legitimacy of the criminal code: they are more fundamentally about which agencies should be responsible for providing the services currently handled by local law enforcement agencies. They are, in other words, about institutional boundaries. Just as improperly bounded institutions can introduce illegiti- macy at the level of the executive, legislature, and judiciary, so too can they render illegitimate the efforts of police agencies.

To this end, Section II develops the idea of a “legitimacy-risk profile” that illuminates which features of institutional arrangements make them tend toward illegitimacy. Sections III and IV draw together common boundaries in policing with legitimacy-risk profiles to explain why certain kinds of policing are especially objectionable and to recommend separa- tions of police power that make policing more legitimate. The resulting framework allows us to make more subtle distinctions about the legiti- macy of political institutions than do theories about whether a state is legitimate overall, highlighting the importance of separations of power all the way down the political hierarchy. Section V concludes the paper.

II. LEGITIMACY-RISK PROFILES

To help determine legitimate separations of power and understand some of the pathologies of policing, I propose legitimacy-risk profiles as a tool. This section will explain what a legitimacy-risk profile is, motivate their use, and detail some of the particular risks relevant to policing.

A legitimacy risk is a characteristic of an entity that decreases the likeli- hood that its political power is permissibly exercised. This is of course quite broad; a police officer who enjoys violating civil liberties is a legiti- macy risk, as is an intemperate president. My focus in this paper shall be on the institutional structures that ameliorate or exacerbate risks to police

2. Minneapolis police officer Derek Chauvin’s killing of George Floyd on May 25, 2020 has led to perhaps the largest political movement in US history, surpassing the civil rights move- ment of the 1950s to 1960s. Larry Buchanan, Quoctrung Bui, and Jugal K. Patel, “Black Lives Matter May Be the Largest Movement in U.S. History,” The New York Times, July 3, 2020.

27 Boundary Policing

legitimacy. Constitutional protections, for example, exist as institutional checks on the legitimacy risks of ill-motivated police officers or presidents. The relevant risks for present purposes are those that can be addressed by appropriate separations of police power.

A legitimacy-risk profile is just the collection of legitimacy risks associ- ated with a political institution. Certain characteristics of an institution are high risk and others are low risk; in addition, the risks can interact and amplify one another. Legitimacy-risk profiles help match institutional arrangements to the particular normative constraints on their goals. This will become clearer after articulating the relevant components of the profiles.

The sense of legitimacy I am relying on is normative. An arrest, for example, is legitimate when it is morally permissible. Legitimacy is con- nected to justice, though less demanding. It does not require perfect jus- tice.3 Legitimacy-risk profiles, as I’ll develop them, are composed of objective standards, combined with an element of democratic authoriza- tion. This contrasts with the descriptive sense of legitimacy found in most police scholarship where police activity is legitimate simply if it is widely accepted as justified by the people being policed.4 Of course, democratic theories of legitimacy blend these together to some extent. But while I will take democratic authorization to be important for under- standing the legitimacy risks of policing, widespread acceptance is not the only relevant factor as descriptive theories of legitimacy might have it.5

To motivate the utility of legitimacy risk in determining separations of power, consider the boundary between police and military. In some ways, the United States is founded on a reaction to illegitimate policing by sol- diers. British soldiers were deployed for law enforcement in the colonies.6

After the revolution, the Bill of Rights prohibited quartering soldiers in pri- vate homes, unreasonable searches and seizures, and required judicial

3. Cf. Buchanan, “Political Legitimacy and Democracy,” 689; John Rawls, Political Liberal- ism (New York: Columbia University Press, 1996), 137.

4. Tom R. Tyler, “Procedural Justice, Legitimacy, and the Effective Rule of Law,” Crime and Justice 30 (2003): 283–358.

5. Descriptive theories of legitimacy are so permissive that they do not even require dem- ocratic authorization.

6. Radley Balko, Rise of the Warrior Cop: The Militarization of America’s Police Forces (New York: Public Affairs, 2013), 213.

28 Philosophy & Public Affairs

approval of warrants. A standing army was controversial in early American political thought; domestic law enforcement by soldiers especially so.

This is not merely a parochial element of American politics. The very first of the “Peelian Principles” (principles apocryphally attributed to the founder of London’s Metropolitan Police, Robert Peel) reads: “To prevent crime and disorder, as an alternative to their repression by military force and severity of legal punishment.” Professional, domestic law enforcement has long been distinguished from military force.7

There is good reason to sharply differentiate military and police activi- ties, and the organizations that conduct them. Good policing requires the exercise of discretion.8 Whereas soldiers can receive frequent and rather specific orders, police officers tend to be subject only to the kind of profes- sional management other professionals are (e.g., performance metrics, contribution to agency-wide goals, etc.). Good police officers seek to deescalate situations and avoid the use of force. Military training encour- ages aggressive, lethal use of force. Police activity is also constrained by a set of civil liberties that do not constrain military activity.

For these reasons, asking standing armies to police the citizenry is highly likely to result in unjustified uses of force. In other words, using the military for policing has extremely high legitimacy risk. One requirement for just and legitimate police power, then, is that domestic policing must be conducted by a separate agency.

If justice and legitimacy require not only separations of legislative and executive power, but further boundaries between kinds of executive power—police and military power—then this will likely entail a require- ment for other boundaries. Some earlier police boundaries were very obvi- ously unjust and illegitimate. Miami once had an entirely self-contained

7. It is true that this is not the case everywhere. Some police functions are provided by a gendarmery—a branch of the military with domestic police power. Still, police functions are often separate from the military functions even when there is not such a sharp institutional boundary. In the Netherlands, some police functions predate the establishment of police departments and were isolated within a specific branch of the military called the Royal Mar- echaussee (Koninklijke Marechaussee). Policing in the Netherlands is also carried out by regional, nonmilitary units; even the Royal Marechaussee answers to a civil, rather than mili- tary leader. What we see in these cases is a different method for achieving the institutional boundary between domestic police and defense military powers.

8. James Q. Wilson, Varieties of Police Behavior: The Management of Law and Order in Eight Communities (Cambridge: Harvard University Press, 1978), 83; David H. Bayley, Police for the Future (New York: Oxford University Press, 1996), 64.

29 Boundary Policing

criminal justice system for black citizens.9 This is an unambiguously ille- gitimate institutional arrangement. But the legitimacy of other arrange- ments is less obvious. Consolidating police departments was thought by many police scholars and policy analysts to be an attractive way to increase the quality of policing. Elinor Ostrom and others found that com- munities prefer smaller, more geographically constrained police depart- ments.10 Legitimacy-risk profiles are a tool for answering normative questions that parallel the ones Ostrom was interested in, ones concerned with the form of police agencies.

One of the virtues of a risk-based framework is that actual political institutions are fallible. The police and the courts will make mistakes, so the standard for legitimacy must accommodate that if any political entity is permitted to exercise political power. It is also occasionally unclear from within the entity when it is making a mistake. This further underscores the need for a framework that provides guidance regarding such a risk.

Another virtue concerns the relationship between individual mistakes and an entity’s overall legitimacy. Enough unjustified action undermines overall legitimacy.11 Yet it is hard to imagine a fully specific account of what it takes for an entity to lose legitimacy. Legitimacy-risk profiles are one such tool because they explain when an entity will tend to act without sufficient justification, and thus tend toward overall illegitimacy, while remaining sensitive to the fallibility of institutions.

Having explained the function legitimacy-risk profiles are to serve and the virtues of the framework, let’s now canvas several of the risks relevant to policing. Risky power is characterized by harshness, high burdens, proactivity, and weak normative mandates. The rest of this section characterizes these legitimacy risks and concludes with a discus- sion of their interactions.

9. W. Marvin Dulaney, Black Police in America (Bloomington: Indiana University Press, 1996), 68.

10. Elinor Ostrom, Roger B. Parks, and Gordon P. Whitaker, “Do We Really Want to Con- solidate Urban Police Forces? A Reappraisal of Some Old Assertions,” Public Administration Review 33, no. 5 (1973): 423–32.

11. This is a simplification for the purposes of the paper. While it seems intuitive that enough illegitimate actions render an agency itself illegitimate, the relationship is compli- cated. I leave the relationship unspecified beyond the claim that high levels of illegitimate activity call into question the agency’s overall legitimacy.

30 Philosophy & Public Affairs

II.A. Harsh and Lenient Power

An institution or enforcement strategy is harsh insofar as it places burdens on those who do not deserve it, and lenient insofar as it fails to place bur- dens on those who do. If a police strategy requires officers to stop and search large numbers of people, most of whom are innocent, the strategy is harsh.

Likewise, a trial system is harsh when it burdens undeserving, innocent defendants, and lenient when it fails to burden deserving, guilty defen- dants. It is widely held that harshness is a more significant problem than lenience.12 In order for a criminal justice system to be legitimate, we must be able to justify the force employed by the system, and false negatives, while regrettable, do not result in the use of force.13 This helps illustrate the distinction drawn above between illegitimate acts of political power and illegitimate political entities. The unavoidable fallibility of a trial sys- tem requires that the system can be legitimate so long as its errors are appropriately patterned.

Policing ought to be generally lenient for similar reasons. After all, in the trial case we’re interested in justifying political power, a consideration equally relevant in policing. One might object here that because the harsh/lenient distinction concerns a distribution of impermissible con- duct, it is not properly understood as a risk factor for impermissible con- duct. Responding to the objection helps to clarify what is meant by harsh policing.

In the case of the trial system, the objection is right; a harsh criminal system constitutes, rather than being a risk for, illegitimacy. In the context of policing, on the other hand, harshness doesn’t automatically entail ille- gitimacy. The reason is that while punishing an innocent person is imper- missible, burdening an innocent person with police power is not always impermissible.

Consider surveillance power: If an officer parks her cruiser in a high crime area to keep an eye on things, she is subjecting everyone in the area to some kind of police power. Most of them will be innocent and will not deserve the burden. This is harsh, though not obviously impermissible. It

12. Cf. Larry Laudan, Truth, Error, and Criminal Law: An Essay in Legal Epistemology (New York: Cambridge University Press, 2006), 63.

13. This is an oversimplification, but I must set aside the problem of discriminatory non- enforcement in this article.

31 Boundary Policing

is, rather, better understood as a risk to legitimacy. In both policing and the trial context, harshness concerns the distribution of burden. But only in the trial context is a harsh distribution necessarily in error.

That the patrol officer’s surveillance is plausibly legitimate should not lead one to think that harshness is not a legitimacy risk. Further reflec- tions on surveillance suggest that harshness itself is a risk. Widespread sur- veillance is objectionable, and many think illegitimate, precisely because it subjects so many innocent people to political power without sufficient rea- son. In the above example, the police power is localized to a crime hot spot, and thus the burden is more easily justified. But if we imagine city- wide surveillance not isolated to crime hot spot locations and times, the increased harshness poses a greater risk to legitimacy.

As a police strategy increases in harshness, the likelihood that it is legit- imate decreases. An intermediate strategy between local hot spot surveil- lance and comprehensive surveillance, such as stop and frisk strategies, is risky because of its harshness. In addition, as we’ll see, even low levels of harshness amplify other legitimacy risks when they interact. We have to consider harshness’s counterpart legitimacy risk: the magnitude of the burden.

II.B. High Burden Power

The magnitude of the burden associated with being subjected to particular instances of political power also helps to determine the legitimacy risk of police activity. The higher the burden, the greater the legitimacy risk. In the case of a criminal trial system, the burden on those found guilty is so high that a harsh distribution undermines overall legitimacy. But the patrol officer’s surveillance is low burden enough that it does not clearly undermine the legitimacy of the strategy.

The magnitude of the burden of political power determines how diffi- cult we ought to make it to subject people to it. A death sentence is extremely high burden. For this reason, the legal path to administering the death penalty is long and expensive (if the path even exists). On the other hand, the burden associated with receiving a speeding ticket is com- paratively low. Given that the burden is rather low, if an innocent motorist is the victim of an accidentally inaccurate radar gun, we take this to be an unfortunate rather than an illegitimate use of force.

32 Philosophy & Public Affairs

We can think of this in terms of proportionality in self-defense or war.14

Just as military force must be proportionate (the amount of force needs to match the expected benefits), so must other instances of political force.15

The higher the burden associated with an instance of political power, the higher the legitimacy risk, because the less likely it is to be proportionate. When the burdens are high, we institute other protections. These include raising the burden of proof in a criminal trial to increase its lenience, cap- ping the penalty associated with criminal activity in sentencing guidelines, or rules prohibiting officers from firing at fleeing suspects.

II.C. Proactive and Reactive Power

A third important component concerns proactive and reactive power. Exercising power proactively involves seeking out problems to address them without being invited to do so by citizens. Exercising power reac- tively, on the other hand, involves responding to existing problems, often at the request of citizens. Reactive police power, more often than not, amounts to third-party defensive force which faces no special justificatory burdens.

Frustrated by the perceived inability of merely reactive policing to deter crime, “problem-oriented policing” has become influential within law enforcement, making policing more proactive.16 Though determining this with any confidence is difficult, proactive policing might in some cases be more effective than reacting to crime and disorder.17 But it also comes with additional legitimacy risks. This is because the police are initiating power. Reactive power, whether political or private, is lower risk. The spe- cial justificatory challenge of preventive warfare, for example, is rooted in the high legitimacy risk of proactive coercive political power.18

14. Helen Frowe, The Ethics of War and Peace: An Introduction (New York: Routledge, 2016), 53.

15. Jake Monaghan, “Legitimate Policing and Professional Norms,” in The Ethics of Polic- ing: New Perspectives on Law Enforcement, ed. Ben Jones and Eduardo Mendieta (New York: New York University Press, 2021).

16. Bayley, Police for the Future, 103; Michael D. Reisig, “Community and Problem- Oriented Policing,” Crime and Justice 39, no. 1 (January 2010): 1–53.

17. Franklin E. Zimring, The City That Became Safe: New York’s Lessons for Urban Crime and Its Control (New York: Oxford University Press, 2012), 142.

18. Frowe, The Ethics of War and Peace: And Introduction, 75–83.

33 Boundary Policing

II.D. Normative Mandates

Consider, finally, strong and weak “normative mandates.” When a law is especially controversial and not widely supported, it is weakly mandated, increasing the legitimacy risk of certain modes of its enforcement. This legitimacy risk is connected to democratic authorization and informs the legitimacy of police agency priorities and enforcement strategies.

Democratic decision making in diverse societies is in part a way to overcome the difficulties of “perspectival diversity.” Perspectival diversity refers to the fundamental moral and descriptive differences in how we understand the world that underlie our policy disputes.19 When elections result in large margins of victory, we can say that either there is low per- spectival diversity on the matter, or that despite the high level of perspec- tival diversity, the various perspectives converge on a political decision. This is why elections with large margins of victory tend to confer stronger justification. The degree of electoral support a candidate receives deter- mines their “manifest normative mandate.”20

Keeping in mind the connection between perspectival diversity and normative mandates allows us to distinguish mere tyranny of the majority from genuine democratic authorization and to see the legitimacy risks associated with prioritizing and aggressively enforcing weakly mandated laws. If a policy is selected by a large majority but split along the lines of relevant political groups, such as when Jim Crow laws were widely supported by white voters, the lack of support across diverse perspectives weakens the law’s normative mandate. Jim Crow laws also show that, in general, the enforcement of laws that will mostly burden the groups who declined to authorize the law are especially weakly authorized, if at all.

When we have significant perspectival diversity and thus a weak nor- mative mandate for a particular policy matter, it is natural to conclude that we should tread lightly. This is close to the core of theories of liberal democratic justification. When we have little diversity and disagreement and thus a strong normative mandate, we have less cause for concern with respect to democratic authorization.

19. Ryan Muldoon, Social Contract Theory for a Diverse World: Beyond Tolerance (New York: Routledge, 2016), 3.

20. Alexander A. Guerrero, “The Paradox of Voting and the Ethics of Political Representation,” Philosophy & Public Affairs 38, no. 3 (Summer 2010): 274.

34 Philosophy & Public Affairs

We need a connection between democratic authorization and legiti- mate policing that takes into account the variations in the strength of justi- fication for laws. The relationship between representatives and voters is underdetermined by an electoral victory.21 The relationship between enforcement strategies and democratic decisions is similarly under- determined. Elections and legislative votes do not tell police agencies how to police. But police agencies have to determine how to allocate their resources and pick their enforcement strategies. In both cases, we can turn to the strength of the relevant normative mandate for guidance. Doing so allows an agency’s decisions to have greater democratic authori- zation rather than being isolated and purely technocratic.

The stronger the normative mandate, the more trustee-like the repre- sentative may be.22 Analogously, the stronger the normative mandate for a law, the greater the strength of authorization for its enforcement. When more strongly authorized, in general, a law’s enforcement can be more aggressive, and agencies are justified in making its enforcement a priority. Proactive enforcement strategies in particular are more likely to be legiti- mate when the law being enforced is more strongly mandated. Whatever our perspective, we will want police to deter and investigate murder, assault, rape, theft, and arson. Laws prohibiting these acts therefore have a strong normative mandate. Indeed, we will want the police to focus on enforcing these laws, and their aggressive enforcement has low legitimacy risk.

On the other hand, when a law is weakly mandated, and especially when the burden of a law’s enforcement falls on the perspective(s) that objected to the law, aggressive, proactive enforcement strategies are more likely to be illegitimate. That police decisions should be influenced by considerations of democratic authorization is implicit in the current wave of “community policing” reform, though it dates back at least to Robert Peel’s conception of policing in which police legitimacy comes from the buy-in of the policed.23 Normative mandates help to understand how that can be achieved. We see here further evidence that risk interactions are particularly important in determining the legitimacy risk of enforcement strategies and agency structure.

21. Ibid., 278. 22. Ibid., 283. 23. Gary Cordner, “Community Policing,” in The Oxford Handbook of Police and Policing,

ed. Michael D. Reisig and Robert J. Kane (New York: Oxford University Press, 2014), 148.

35 Boundary Policing

II.E. Risk Interactions

In isolation, these risks might not pose significant risks to legitimacy. Taken together, they interact to amplify (or mitigate) total legitimacy risk. Recall the patrol example discussed earlier. The magnitude of the burdens of power interacts with the distribution of burdens to determine legitimacy risk. Harsh, but low burden policing, such as patrol officer surveillance of a crime hot spot, is rather low legitimacy risk. Increasing the magnitude of the burden increases the risk; as the magnitude increases, we need to reduce the strategy’s harshness (hence constitutional restrictions on stop- ping and searching people without probable cause or a warrant). Gener- ally, political systems that subject citizens to high burden political power had better be lenient, whereas harsh but low burden political power is more likely to be legitimate.

Consider again the connection between normative mandates and pro- active power. Reactive policing has been characterized by political scien- tist David Bayley as making policing more democratic by giving citizens a role in how resources are allocated and how police power is focused. As Bayley puts it, technological advancements like the telephone meant that

demand for police services no longer emanated solely from the state but was distributed across a multitude of individuals. Demand was dis- aggregated, and the police developed a new set of masters. The police became democratically responsible in terms of their day-to-day operations.24

Proactive power for this reason often, though not always, entails a wea- ker normative mandate. The proactive enforcement of weakly mandated laws is thus especially risky; its authorization is doubly weak.

Different police strategies will be characterized to different extents by these risks, and the risks all interact. Let us say, then, that legitimacy risk increases when political power is harsh, high burden, proactive, and high diversity/weakly mandated. Conversely, when political power is lenient, low burden, reactive, and low diversity/strongly mandated, it tends to have a low legitimacy risk. These risk factors are related, though each can individually affect instances of political power. In other words, it is

24. Bayley, Police for the Future, 120.

36 Philosophy & Public Affairs

possible for power to be both reactive and conducted against a back- ground of high perspectival diversity, and these simultaneously increase and decrease the legitimacy risk.

As a disclaimer, it is worth emphasizing again that these profiles track risk; it is possible for a high-risk police unit to act legitimately and have overall legitimacy. But given the uncertainties of realized institutions and political power, the framework encourages policymakers (and voters) to opt for low-risk police agencies and strategies.

How likely we think there are to be unjustified uses of power largely determines how many safeguards we will want to erect. If elected officials were always well motivated to respect civil liberties, we might not want a system that includes judicial review. If we imagined the most ideal sol- diers, perhaps they would be talented and well motivated enough to engage in both military action and domestic policing concurrently and in the same organization. Because actual soldiers are not ideal soldiers, and because changing one’s rules of engagement drastically and instantly is incredibly difficult, we mitigate legitimacy risk institutionally. Importantly, different kinds of policing have different legitimacy-risk profiles just as military and police power have different profiles. Hence, the need for a more fine-grained approach to separations of political power.

III. VARIETIES OF POLICING

Having laid out the ingredients of legitimacy-risk profiles, I now want to consider some varieties of policing to illustrate the application of the framework. What we will see in this section is that recent developments in policing have tended to blur together police goals, thereby blending— rather than isolating—high legitimacy-risk characteristics. This will put us in a position to see which separations of power the legitimacy-risk profile framework supports in the following section.

Modern police agencies have three main roles: administrators, patrol officers, and detectives. Originally, most early policing was aimed at order maintenance; crime investigation and law enforcement was reserved for private detectives.25 The difference between the patrol officer and the

25. There is an important exception: slave patrols in the antebellum south were not maintaining order, but rather were enforcing (obviously unjust and illegitimate) laws.

37 Boundary Policing

detective roughly tracks the distinction between order maintenance and law enforcement.

By keeping their eyes on the street and rapidly responding to crimes in progress or calls for service, patrol officers should prevent crime from hap- pening and mediate disputes that do not rise to the level of litigation. The patrol officer could conceivably do all of her work without making any arrests. In fact, only a minority of what they do is enforcing laws and mak- ing arrests. They aim to “handle the situation” or to provide “authoritative intervention.”26 That is not to say that patrol officers never enforce law: when they see or respond to criminal activity in progress, they are expected to enforce law by making an arrest. But whereas enforcing law is the primary aim of the detective, it is only a secondary aim of the patrol officer.

Order maintenance patrol policing therefore tends to be low burden and characterized by wide discretion. The ill-defined and high perspectival diversity nature of the issues patrol officers are called to handle (e.g., disorderly conduct) are offset by the low burden nature of patrol policing. Given the costs of going to trial and conviction, the detective’s work tends to be high burden and constrained by procedural rights. These procedural rights, combined with the clearly defined and typically strongly mandated nature of the laws enforced by detectives, compensate for the risk-increasing nature of the burdens of this kind of policing.

Historically, the patrol officer’s and the detective’s work are reactive, and this serves to further compensate for the relevant legitimacy risks (the high burden associated with felony prosecution and the often- controversial nature of what counts as order). Rapid response to calls for service is of course reactive. Foot patrol or motor patrol is proactive only in the sense that the officer is patrolling areas to deter crime and to be able to quickly react to problems. Similarly, homicide, violent, and prop- erty crime detectives rarely engage in work that someone has not specifi- cally asked them to engage in. They react to reported crimes and proceed based on clearly defined legal authority to investigate, prosecute, and incarcerate serious wrongs.

As an illustration, consider the classic image of a police officer: Officer Friendly walking a “beat,” waving or smiling at familiar faces, and provid- ing reassurance that if something goes wrong, he will be there to

26. Wilson, Varieties of Police Behavior, 31; Bayley, Police for the Future, 34.

38 Philosophy & Public Affairs

intervene. If everything goes according to plan, the officer never needs to intervene. Those who would misbehave know that someone is standing by to stop them if they get into trouble. The officer is creating a secure and ordered environment simply by making the rounds. Such an officer is wielding a kind of counterfactual political power: they will intervene if called for. Typically, however, they do not intervene, and the power involved in occasionally coming under this officer’s “surveillance” is rather easily justified.

But suppose a man stumbles out of a bar having had too much to drink and begins bothering passersby; the officer intervenes to send the man on his way. The kind of power involved here seemingly requires no special justification, as a courageous bystander could sternly ask the man to keep moving, and no one would find this objectionable. This use of power is relatively widely accepted. Only the drunk would be bothered by this exer- cise of power, and then only while still drunk. When an arrest is not the go-to tool, the officer’s power is low burden. And better still, those sub- jected to arrest power are likely to be guilty. So Officer Friendly’s power is lenient. On most metrics of legitimacy risk, the classic foot-patrol officer’s is low.

Unfortunately, this kind of police officer has always been rare and is increasingly a relic of the past. Patrol officers often abuse their discretion- ary power to make disorderly conduct arrests for arbitrary—or worse, discriminatory—ends.27 “Officer Friendly” is a term now used sardonically to refer to abusive patrol officers. Most citizens want a very fast response time to calls for police, so agencies have responded by moving the patrol officer into an SUV. This separates officer from citizen, depriving them of an important source of local knowledge.

There are other reasons that Officer Friendly is a relic and riskier kinds of policing are more common. Empirical investigation of patrol policing in the 1970s suggested that this kind of reactive patrol does little to deter crime.28 By the 1980s, there was general skepticism about the effectiveness of police agencies in reducing crime, with particular skepticism about

27. James Forman, Jr., Locking Up Our Own: Crime and Punishment in Black America (New York: Farrar, Straus and Giroux, 2017), 94; Sarah A. Seo, Policing the Open Road: How Cars Transformed American Freedom (Cambridge: Harvard University Press, 2019), 169, 183, 256.

28. George L. Kelling, Tony Pate, Duane Dieckman, and Charles E. Brown, The Kansas City Preventive Patrol Experiment (Washington, DC: National Police Foundation, 1974).

39 Boundary Policing

reactive patrol. This motivated George Kelling and James Wilson to articu- late the “broken windows” approach to policing which encouraged some aggressive police tactics. Many saw this as an improvement on overly pas- sive patrol.29 Increased federal funding and asset forfeiture policies encouraged multijurisdictional narcotics task forces and changes in local police priorities.30 These changes contributed to the erosion of the bound- ary between patrol and detective work.

Under Mayor Giuliani and New York City Police (NYPD) Commissioner William Bratton, for example, low-level arrests spiked. Partially motivated by the adoption of COMPSTAT, priorities in patrol work changed. Patrol officers were to focus on making street stops and arrests; foot patrol and motor patrol became criminal patrol. This persisted after Giuliani left office with the stop, question, and frisk (SQF) strategy. Plenty of ink has been spilled over the development and spread of broken windows policing, SQF, and related theories of the growth, spread, and deterrence of crime. The important point in the present context is that transforming the reactive patrol officer into a proactive stopping, questioning, and frisking officer, or into one concerned with making arrests for “quality of life” violations, entails a substantial change into a different kind of legitimacy-risk profile.

The SQF strategy employed by the NYPD was eventually found uncon- stitutional. Nevertheless, this kind of proactive policing is now widespread, as is the use of COMPSTAT or similar programs.31 The LAPD’s Metropoli- tan Division doubled in size in 2015, with many of the officers going to squads that engage in proactive policing: traffic stops aimed at discovering guns and drugs. Referred to as “stop and frisk in a car,” the LAPD, like the NYPD, stop black motorists at a disproportionate rate. Many of the people stopped are stopped repeatedly, and while the rate of traffic stops went up as much as 14 times, felony arrests merely tripled.32

SQF tactics rely on proactively stopping large numbers of people to sea- rch for evidence of certain crimes. Most of the people stopped are

29. George L. Kelling and James Q. Wilson, “Broken Windows: The Police and Neighbor- hood Safety,” The Atlantic Monthly, 1982; Bayley, Police for the Future, 103.

30. Peter J. Boettke, Jayme S. Lemke, and Liya Palagashvili, “Re-Evaluating Community Policing in a Polycentric System,” Journal of Institutional Economics 12, no. 2 (June 2016): 305–25.

31. Peter Moskos, Cop in the Hood: My Year Policing Baltimore’s Eastern District (Princeton: Princeton University Press, 2009), 111–57.

32. Cindy Chang and Ben Poston, “‘Stop-and-Frisk in a Car’: Elite LAPD Unit Dispropor- tionately Stopped Black Drivers, Data Show,” Los Angeles Times, January 24, 2019.

40 Philosophy & Public Affairs

innocent.33 Some of the crimes targeted are controversial (like drug pos- session). The tactic itself is controversial. Repeated stops can be costly in terms of individual well-being.34 In other words, the tactic is not only weakly mandated, it is also harsh and high burden. The tactic’s proactivity further amplifies its risk.

When order maintenance patrol turns into criminal patrol, the burdens increase in another way. When stops result in arrests, the misdemeanor system becomes congested. This carries with it a variety of ill effects. As Alexandra Natapoff documents, the misdemeanor system becomes more punitive when it is congested, including overburdening public defenders and increasing the incentive for prosecutors to expedite cases as much as possible. This results in lower quality pre-trial screening of cases, reliance on coercive plea deals, and excessive deference to the arresting officer’s testimony. This produces not only punishment without crime, but punish- ment that doesn’t match the crime.35

Criminal patrol also bleeds one police goal into another, suppressing the original goal. The patrol officer who was supposed to mediate disputes and deter crime is now trying to make arrests and enforce the law. The traffic officer who was supposed to produce safe, ordered use of transpor- tation infrastructure now uses their authority in service of a very different kind of policing. In a how-to manual for “aggressive criminal patrol,” offi- cers learn how to turn patrol into something that allows them to “pull off make-a-difference busts.” These officers “regard themselves as first-line criminal investigators . . . total-commitment law enforcement officers.”36

The result is not only the familiar problems associated with SQF and “driving while black,” but an abandonment of a longstanding service of reactive, order maintenance patrol in favor of the much riskier proactive criminal patrol. The goal of stopping a vehicle carrying drugs substantially conflicts with the goal of maintaining safe roads and transforms a low risk kind of policing into a much riskier kind.

33. Robert Apel, “On the Deterrent Effect of Stop, Question, and Frisk Stop, Question, and Frisk Practices,” Criminology and Public Policy 15, no. 1 (2016): 57–66.

34. Amanda Geller et al., “Aggressive Policing and the Mental Health of Young Urban Men,” American Journal of Public Health 104, no. 12 (October 16, 2014): 2321–7.

35. Alexandra Natapoff, Punishment Without Crime: How Our Massive Misdemeanor Sys- tem Traps the Innocent and Makes America More Unequal (New York: Basic Books, 2018).

36. Charles Remsberg, Tactics for Criminal Patrol: Vehicle Stops, Drug Discovery and Offi- cer Survival (Northbrook: Calibre Press, 1995), 9.

41 Boundary Policing

Consider now the changes in the detective’s role. The homicide unit is low legitimacy risk; it is reactive and strongly mandated. Though an arrest for homicide is very high burden, the procedural constraints help to ensure that the police power is not harsh. The same cannot be said for other kinds of detectives.

Alongside a more aggressive approach to patrol work, there was a rise in “vice” policing. In 1972, the Office of Drug Abuse Law Enforcement began conducting drug raids following Nixon’s declaration that drugs are “public enemy number one” the year prior. In 1973, ODALE merged with the Bureau of Narcotics and Dangerous Drugs to form the Drug Enforce- ment Administration. Here too we see a rise in proactive police work. The vice or narcotics detective is perhaps uniquely proactive.37 “Buy-bust” “sting” operations, popular with narcotics and vice policing, are proactive. They set up the conditions for criminal activity, and when the encouraged crime is committed, detectives make an arrest.38

Vice and narcotics law enforcement is differentiated from the homicide or violent crime detective in terms of initiation and authorization:

When a violent crime or a robbery or a trespass occurs, someone usu- ally calls the police. There is a clear victim and perpetrator. Someone is hurt or harmed in some way and wants the offender punished. But with drug crime, neither the purchaser of the drugs nor the seller has any incentive to contact law enforcement. It is consensual activity. Equally important, it is popular. The clear majority of Americans of all races have violated drug laws in their lifetime.39

Similarly, a detective working in a vice unit might get information that a certain area sees high rates of sex work, and on the basis that a crime is likely to be committed, stake out part of the area. They might pretend to be a customer or sex worker to induce criminal behavior in the potential criminal and make an arrest. But as with drug enforcement, the laws being enforced are weakly mandated, and typically, those involved would like to be left alone. Vice work is therefore high burden, typically proactive, high diversity, and weakly mandated, though unlike the move to criminal patrol, perhaps not harsh.

37. Wilson, Varieties of Police Behavior, 9. 38. They also often involve law enforcement participation in illegal activity. This generates

legitimacy-risks of its own. See Luke William Hunt, The Retrieval of Liberalism in Policing (New York: Oxford University Press, 2019), 172.

39. Michelle Alexander, The New Jim Crow (New York: The New Press, 2012), 104.

42 Philosophy & Public Affairs

It is possible that the drug enforement administration (DEA) or local narcotics divisions are behaving legitimately. In cases where police are reacting to complaints from citizens, there is a genuine conflict of interests that can be difficult to settle; determining which course of actions are legitimate is highly difficult in these contexts. Nevertheless, in most cases, the combination between high perspectival diversity, weakly mandated, high burden, and proactive law enforcement makes their work highly risky—perhaps the riskiest form of policing in the United States. One vir- tue of the legitimacy-risk framework is that it enables such a comparative evaluation without relying on foundational arguments against the legiti- macy of narcotics or vice enforcement.40

It is worth noting that this kind of concern is not new. In the 1920s, influ- ential police chief August Vollmer complained in a sardonic law review arti- cle about the public’s expectation that police enforce vice and traffic law.41

The automobile was new at the time, and the new traffic laws were increas- ingly used to target alcohol smugglers during prohibition. He saw vice enforcement as hopelessly ineffective and traffic enforcement as a distrac- tion from real police work. The only things that have changed are the vices being prohibited and the now-widespread acceptance of police enforce- ment of traffic law. Vollmer was right that these activities detracted from more important police goals. Legitimacy-risk profiles help us to see how they produced a major shift in policing, and not only in its focus; policing has become more proactive, harsher, higher burden, as well as focused on weakly mandated laws, and employing weakly mandated tactics.

IV. INSTITUTIONAL BOUNDARIES

Legitimacy-risk profiles provide a tool for understanding how some prob- lems of modern policing are rooted in insufficient boundaries between

40. There are of course many other kinds of policing that can be evaluated using the legitimacy-risk profile framework. Special Weapons and Tactics (SWAT) teams, for example, are now used primarily for serving drug warrants, making it weakly mandated, proactive, and extraordinarily high burden. Specialized units for policing protests will be legitimate to the extent that they are low burden, reactive, and strongly mandated. Federal agencies that focus on human trafficking, for example, might be proactive and high burden, though very strongly mandated and lenient. Because it should be clear how the framework applies, I leave these cases unanalyzed here.

41. August Vollmer, “Vice and Traffic-Police Handicaps,” Southern California Law Review 1, no. 4 (1928): 7.

43 Boundary Policing

varieties of policing. In this section, I’ll argue that the different legitimacy- risk profiles of various kinds of policing recommend certain institutional separations of police power. Some of the redrawn boundaries would undo the changes discussed in the previous section, but I’m not motivated by nostalgia for early twentieth-century policing. Legitimacy-risk profiles pro- vide guidance for improving the legitimacy and justice of policing beyond simply undoing the mistakes of the past. In particular, boundaries moti- vated by the legitimacy-risk framework prevent the combination of high legitimacy-risk factors, institutionally separate the enforcement of weakly mandated laws, and allow agencies to focus on particular police goals to prevent certain goals from crowding out others. Drawing new institutional boundaries reduces the risk of losing legitimacy and increases opportuni- ties for developing legitimacy. Because illegitimate police power is unjust, reducing legitimacy risks will also tend to reduce unjust policing.

The relevant boundaries are jurisdictional: geographical, professional, and legal. Geographic boundaries of course concern the area an agency or division should have authority over. Professional boundaries concern the methods and goals of the agency, including the tools at their disposal. Those tools include the areas of the civil or criminal legal code the agency enforces. The framework developed in this paper shows that legitimacy requires paying careful attention to each kind of boundary.

IV.A. Geographic Boundaries Strengthen Normative Mandates

Only some policing is characterized by high perspectival diversity and so weakly mandated. Violent, sex, and property crime law enforcement are strongly mandated. Thus, there is no reason in principle to geographically constrain the jurisdiction of detectives working on these matters; a state or even nationwide police force focused on investigating and making arrests for violent and property crimes has a reasonably low legitimacy-risk pro- file. Considerations of efficiency (in terms of clearing cases) should be weighed most heavily when thinking about the geographic boundaries for these kinds of agencies.

Order-maintenance policing, on the other hand, is characterized by rel- atively high perspectival diversity. Only the sticks-in-the-mud are bothered by open alcohol containers in New Orleans’ French Quarter, whereas open alcohol containers near an elementary school gives cause for con- cern. Similarly, one would have to be among the most obnoxious people

44 Philosophy & Public Affairs

the world has ever seen to complain about street musicians on French- man Street in the evening, whereas one would have to be comparably obnoxious to busk on a street corner in a suburban neighborhood at night.

What counts as order in a public space depends on a number of partic- ular facts, and those particular facts can change from block to block and neighborhood to neighborhood. Thus, geographically constraining the jurisdiction of order-maintenance police forces is the only way to reduce the legitimacy risks associated with underpolicing or overpolicing an area and to engage in strongly mandated policing. A statewide force is unlikely to produce legitimate order-maintenance policing because it is unlikely to react appropriately to a community’s needs. Drawing district boundaries within an agency might help, though there is an ineliminable risk that administrators with authority over many districts will pressure officers on the streets to pursue standardized goals.

Right now, it is common for departments to divide patrol into geo- graphic divisions and for professional specialization to occur outside the patrol division. There is room for professional specialization within these divisions that can lower legitimacy-risk profiles.

IV.B. Professional Boundaries Isolate Legitimacy Risk

The legitimacy risks of military power require professional, institutional boundaries. The same goes for different kinds of policing. But as discussed above, the proactive turn in law enforcement has spread risky strategies to most other facets of law enforcement.

Looking at one of the mechanisms by which classic patrol was trans- formed into criminal patrol, Operation Pipeline, will clarify the importance of certain institutional boundaries. The DEA’s Operation Pipeline was a deliberate attempt to blur the institutional boundaries between federal drug law enforcement and local patrol.42 The DEA convinced highway patrol to target suspected drug trafficking (rather than mere moving viola- tions) in part by providing the training for free. Asset forfeiture rules allowed sharing between the DEA and local agencies which provided additional motivation for criminal patrol.

This led to instrumental (or pretextual) enforcement of traffic laws. This changed the focus of traffic officers from maintaining traffic safety to drug

42. Seo, Policing the Open Road, 256.

45 Boundary Policing

or gun interdiction (or better yet—cash to be seized and kept by the department). These kinds of stops are highly unlikely to result in relevant charges because it is not clear by looking at a vehicle that it will contain contraband, and only a small minority of them do. The result is that the kind of behavior we might be interested in deterring goes largely ignored, while large numbers of innocent (often black) motorists are repeatedly stopped and harassed. The DEA’s legitimacy risks have infected local law enforcement, making it harsher, higher burden, proactive, and weakly mandated. Institutional boundaries require not just separate agencies, but financial rules that maintain the boundaries.

If the drug war is an infection, its harmful delegitimizing effects can be ameliorated and contained. One way to isolate legitimacy risks is to insti- tutionally separate the agents enforcing criminal law from the agents enforcing traffic laws. Narcotics divisions should not only be separated from the traffic division (as is currently the case), but perhaps traffic offi- cers should have to call for a narcotics detective to conduct a vehicle sea- rch, just as homicide detectives have to be called in to conduct an investigation. This will surely make narcotics investigations more cumber- some. Given its high legitimacy risk, that is a feature and not a bug.43

IV.C. Specialization Changes Selection Effects

In light of the seemingly widespread endorsement of criminal patrol within the profession, it is also necessary to have officers interested in the kind of policing their unit is primarily tasked with. Informal norms are an important component of professional behavior. Fortunately, institutional boundaries are useful here as well. Institutional boundaries can foster goal-oriented specialization, changing the selection effects of employment within an agency. This in turn can have a substantial effect on the way an agency behaves.

The DEA may be partially responsible for eroding the boundary between traffic enforcement and drug interdiction, but it is not entirely responsible for the move to criminal patrol. In criticizing the aggressive nature of patrol policing, former officer Peter Moskos remarks that:

43. Legitimacy-risk profiles can also recommend resource allocations within an agency or across agencies. A local department will want to allocate more resources to the homicide unit than the vice unit; the federal government will want to allocate resources away from the DEA and toward the FBI’s human trafficking task forces.

46 Philosophy & Public Affairs

Just as a culture of force gave way to a culture of arrest and “zero tolerance,” the arrest culture needs to evolve into something better: a culture of crime prevention, problem solving, and police discretion. As laws against jaywalking do not fill the prisons with jaywalkers, laws against drug use did not and need not fill our jails with drug users. . . . To paraphrase a slogan of the National Rifle Association: drug laws don’t arrest people, police do.44

Specialization can be a useful tool in changing the culture of arrest, replacing Remsberg’s “first line criminal investigators” with (much mal- igned by the police) “social workers with guns,” thereby changing the legitimacy-risk profile of the patrol division.

The typical urban police department has specialized investigative units, though “the undifferentiated patrol division gets everything else.”45 The undifferentiated patrol division is not well suited to focus on the incredibly complex and normatively difficult issues of order maintenance. And as we become increasingly aware of the ham-fisted way that patrol officers han- dle those with mental (or in some cases physical) illness, and the atten- dant cost to their life and/or well-being, this seems increasingly hard to justify. Of course, Wilson wrote this in the 1960s and policing has under- gone extensive change during this time. Nevertheless, while specialization has grown, law enforcement activities are highly specialized and order maintenance activities are less so. Splitting the patrol function and the detective function into different agencies, and then specializing the patrol division, would help to enable the kind of specialization within patrol that seems badly needed.

One mechanism for changing the nature of patrol policing depends on changing the hierarchy in police agencies. Nearly all police officers now start out in patrol. If they do well by the standards of the administration, they are promoted to the rank of detective. If they continue to do well, they might have a chance at promotion to administration. There are sev- eral noteworthy results.

One is that detectives enjoy a kind of prestige that patrol officers lack, in addition to greater pay and privileges. No one is promoted to patrol. This means that most of the people who seek promotion want to look like

44. Moskos, Cop in the Hood, 185. 45. Wilson, Varieties of Police Behavior, 69.

47 Boundary Policing

effective law enforcers. Given this career ladder, the metrics for evaluating patrol work are likely skewed toward identifying skills that make one a good law enforcer. So police departments end up selecting for people who want to arrest (and help punish) the bad guys, rather than people who want to provide the much-needed professional support in coordinating on the use of public space and the other various tasks patrol officers engage in that look much more like social work than stereotypical police work.

Punishment of some sort is plausibly a necessary condition for justice, so I don’t mean to imply that having this as a career goal is vicious. The point is that individuals with punitive ambitions are suited for law enforcement, and less well suited for patrol. To reduce legitimacy risks, we want police agencies to select for people with helping and healing goals, not punitive goals.46

Furthermore, specializing the separate patrol agency would allow offi- cers who are interested in helping, say, victims of domestic abuse to focus on that kind of police work. It would also allow more specialized training, so that patrol officers who do wellness checks are better suited to handle those calls. As things are now, police training ends up looking like a zero- sum game with, for example, firearm training reducing resources for men- tal health training.

More radical versions of this proposal include removing traffic enforce- ment and wellness checks from police departments entirely and turning them over to civil servants without police powers (like parking enforce- ment officers) and medical professionals. I won’t defend a specific version of this boundary redrawing proposal; I’ll just note that legitimacy-risk pro- files are consistent with more and less ambitious reform ideas.

IV.D. Specialization Increases Democratic Authorization

Lastly, in addition to drawing geographic boundaries to make agencies more sensitive to community needs and drawing professional boundaries to allow for specialization that remakes agency goals, selection effects, and training standards, we can also turn to specialization to increase demo- cratic authorization. More specialization allows for more targeted

46. An additional benefit of separating the detectives from the patrol officers is that inves- tigations of police brutality might be more impartial.

48 Philosophy & Public Affairs

democratic input in two ways: political approval of agency goals and methods, and democratic control over their funding.

Now that killings by police are a national political issue, police officers take themselves to be under fire and disrespected. But in many cases, the targets are aggressive and exploitative patrol officers. By separating patrol from detectives, protesters can target certain agencies without being accused of general “anticop” sentiment. Members of one agency could plausibly support certain protests without feeling like they are also being “anticop.” Legislators could campaign on remaking or reducing funding for badly behaved agencies without seeming to disregard the goals of policing. Politicians and activists today who want to eliminate the U.S. Immigration and Customs Enforcment (ICE) agency or the DEA are targeting certain kinds of objectionable policing; this only happens when different kinds of policing are separated institutionally. Given the drastic separation between a vote for a legislator and the power exercised by one officer, there is a need to reform political institutions such that they can authorize or deauthorize police power. Police specialization can help clar- ify the strength of the normative mandate for particular agencies and their strategies.

V. CONCLUSIONS

I have argued that the ills of contemporary policing are essentially con- nected to the poorly drawn boundaries of police power. Effective police reform will require further separations of power. Legitimacy-risk profiles composed of a measure of harshness, the magnitude of burden imposed by police activity, proactivity, and normative mandates, are a useful con- ceptual tool for understanding our current problems and for identifying legitimacy-enhancing separations of power. The main institutional bound- aries recommended by applying legitimacy-risk profiles are those that iso- late positive legitimacy-risk interactions, strengthen normative mandates, clarify which aspects of policing are weakly mandated, and match selec- tion effects and training standards to the agency’s proper goals. Practically speaking, this means increasing the institutional boundaries between order maintenance and law enforcement functions to prevent the goals of apprehending and punishing criminals from overwhelming the goals of protecting and serving people, strengthening the boundary between traffic enforcement and other kinds of law enforcement, and specializing the

49 Boundary Policing

patrol division. It also means separating particularly controversial kinds of policing to allow for greater democratic authorization. Thinking in terms of legitimacy-risk profiles suggests that we have so far paid insufficient attention to the relationship between legitimacy and separations of power.

50 Philosophy & Public Affairs